2

A few days ago I wanted to use float128 from Boost multiprecision but it turns out only ICC and GCC support float128:

CMake, Boost multiprecision float128 and Clang "fatal error: 'quadmath.h' file not found"

I am looking for a performant 128-bit Decimal library for C++ (I know float128 is different).

Does this also mean it's impossible to use the Intel Floating Point Math Library with Clang?

https://www.intel.com/content/www/us/en/developer/articles/tool/intel-decimal-floating-point-math-library.html

The README mentions they tested using GCC, but doesn't explicitly state only ICC or GCC must be used.

rare77
  • 297
  • 6
  • 1
    C++23 added support for 128-bit floats: https://en.cppreference.com/w/cpp/types/floating-point, so it's now part of the language, but clang doesn't have compiler support for this feature yet. – Jan Schultke Jun 14 '23 at 23:27
  • @JanSchultke You've closed my question, informed me about floats (thanks) but my question is really about the Intel fixed decimal library..... ? And it's not asking for recommendations, it's asking if anything exists. – rare77 Jun 15 '23 at 00:07
  • 1
    "I am looking for a performant 128-bit Decimal library for C++" is asking for a recommendation. You could always try the Intel library you mentioned with your compiler of choice and see if it works. If not you might ask about how to get it to work and explain the issues you had that prevented it. – Retired Ninja Jun 15 '23 at 00:14
  • recommendations are off-topic. So if you want to reopen this you need to edit it to ask something like *how to use 128-bit float in Clang* – phuclv Jun 15 '23 at 01:41
  • 2
    @JanSchultke they're **optional** and only available when the macros are defined, specifically [`__STDCPP_FLOAT128_T__` for `std::float128_t`](https://en.cppreference.com/w/cpp/header/stdfloat) so there's no requirement for an implementation to support it – phuclv Jun 15 '23 at 01:47

0 Answers0